Highly Available Queue Prototype
--------------------------------

Original implementation by Henry Robinson.
* http://github.com/henryr/pyzk-recipes 

I've changed the code to retry the operation on ConnectionLoss.

How to play with the demo?
--------------------------

$ cd test_5_trunk
$ ./start.sh
$ python fail.py 

in another shell do:

$ ./top.sh

As you can see the cluster is pretty unstable. 
fail.py kills and restarts one random node at a time. 
Even if this happens the application should be able
to easily survive server failures. 

Start the queue:

$ python queue.py 

You should see multiple "Connected" messages.